194 Int. J. Advanced Mechatronic Systems, Vol. 1, No.

3, 2009

Design of low-cost unmanned underwater vehicle for

shallow waters

W.H. Wang* and X.Q. Chen

Dept. of Mechanical Engineering,
University of Canterbury,
Private Bag 4800, Christchurch 8140, New Zealand
E-mail: [email protected] E-mail: [email protected]
*Corresponding author

A. Marburg
Geospatial Research Centre (NZ) Ltd,
Private Bag 4800, Christchurch 8140, New Zealand
E-mail: [email protected]

J.G. Chase and C.E. Hann

Dept. of Mechanical Engineering,
University of Canterbury,
Private Bag 4800, Christchurch 8140, New Zealand
E-mail: [email protected] E-mail: [email protected]

Abstract: Unmanned underwater vehicles (UUVs) have received worldwide attention and been
widely used in various applications. In this paper, a recently developed low cost UUV prototype
at the University of Canterbury is introduced, which is designed specifically for shallow water
tasks, especially for inspecting and cleaning sea chests of ships for bio-security purpose. The
main hull of the UUV is made of polyvinyl chloride (PVC), with a 400 mm diameter and
800 mm length. External frames mount two horizontal propellers, four vertical thrusters and
power is derived from onboard batteries. The maximum thrust force of up to 10 kg that is
provided by the propellers can generate a forward/backward speed of up to 1.4 m/s for the 112 kg
UUV. The vertical thrusters provide depth control with a max thrust force of 20 kg. The UUV is
equipped with a range of sensors capable of sensing its instantaneous temperature, depth, attitude
and surrounding environment. Costing less than US$10,000 for a prototype, it provides an
excellent platform for further underwater vehicle development targeting shallow water tasks with
a working depth up to 20m.

Keywords: unmanned underwater vehicle; UUV; low cost; hull design; shallow water
inspection.

Reference to this paper should be made as follows: Wang, W.H., Chen, X.Q., Marburg, A.,
Chase, J.G. and Hann, C.E. (2009) ‘Design of low-cost unmanned underwater vehicle for shallow
waters’, Int. J. Advanced Mechatronic Systems, Vol. 1, No. 3, pp.194–202.

Biographical notes: Dr. Wenhui Wang is Lecturer of Mechanical Engineering Department. His
research areas include computer vision, micromanipulation, bio-device and instrumentations,
visual servo control, and manufacturing process monitoring. He has published over 35 peer
viewed conference and journal papers, holds three US patents, and is on the Editorial Board of
three journals in mechatronics.

Dr. Xiao Qi Chen is Associate Professor with Department of Mechanical Engineering, University
of Canterbury, and the Director for the Mechatronics Engineering Programme. He has published
over 100 refereed journal and conference papers and is Associate Editor for several known
international journals.

Mr. Aaron Marburg is a Research Scientist at the Geospatial Research Centre NZ. He is
interested in innovative techniques for scientific sampling including autonomous vehicles and
wireless sensor networks.

Dr. J. Geoffrey Chase is Professor with Department of Mechanical Engineering, University of

Canterbury. His research interests include: biomedical systems and devices, structural dynamics
and vibrations, dynamic systems modelling, automatic control, MEMS and micro-fabrication,

Copyright © 2009 Inderscience Enterprises Ltd.

 Design of low-cost unmanned underwater vehicle for shallow waters 195

and embedded systems. He has published over 320 refereed journal and conference papers and is
an inventor on over ten patents. He is on the Editorial Board of three journals in
bio-engineering and medicine.

Dr Christopher E. Hann is a Sir Charles Hercus Health Research Fellow and Senior Research
Associate. His research areas include computational and physiological modelling for biomedical
systems, model-based diagnostic and therapeutics for critical care, physiological parameter
identification and the application of image processing and computer vision in Biomedicine.
He has published over 115 refereed journal and conference papers, is an inventor on several
patents and is on the Editorial Board for an Open Bentham journal in medical informatics.

1 Introduction Grates on the outside of the chests prevent large organisms

from being entrained in the water but many smaller
Unmanned underwater vehicles (UUVs) offer many
organisms [Figure 1(b)] survive in the sea chests and are
advantages for performing difficult tasks submerged in
transported around the world creating a bio-security risk.
water. Although many names are given to these vehicles,
The New Zealand government has placed a high priority on
such as remotely operated vehicles (ROVs), submersible
the development of systems and tools to protect native flora
devices or remote controlled submarines, to name just a few,
and fauna against invasion by unwanted foreign organisms.
the fundamental task for these devices is fairly well defined.
They operate in conditions and perform task that humans are Figure 1 UUV inspecting and cleaning sea chest of ships, (a) the
not able to do efficiently or at all (Smallwood and diagram of the UUV working on the sea chest of the
Whitcomb, 2004; Horgan and Toal, 2006; Caccia, 2006). ship (b) a range of foreign invaders hiding in the sea
First developed in the 1960s, development was driven chest (see online version for colours)
by the demand from the US Navy (Wernli, 2001), which
required them to perform deep sea rescue and salvage
operations. Since then, other sectors have realised the
potential of such devices for all manner of tasks. The first of
these was the oil and gas industry. These companies
employed UUVs to reinforce in the development of off
shore oil fields (Williams, 2004). In the 1980s, UUVs came
into a new era as they were able to operate at depths well
below commercial diver limits. Falling oil prices and a
global recession resulted in a stagnant period in terms of
UUV development in the mid 1980s. Since then however,
UUVs have been developing at a fast rate (Smallwood et al.,
1999; Griffiths and Edwards, 2003).
UUVs are now being used in a wide range of
applications, such as locating historic ship wrecks like the
Titanic (Ballard, 1987), mapping the sea floor (Tivey et al., To optimise the knowledge of, and reaction to, this threat,
1998). More mundane applications consist of object the first task is to inspect the sea chests and collect
detection (Kondoa and Ura, 2004), securing harbours, information about the invaders. Currently, divers are sent to
searching for seamines (Willcox et al., 2001) and, most do the job, which has inherent problems, including:
recently, in scientific applications (Curtin and Bellingham,
1 high cost
2001; Rife and Rock, 2002; Lygouras et al., 1998).
As more than half of our oceans are deeper then 3 km, 2 unavailability of suitably trained personnel for the
one direction of the UUV developments is to explore deep number of ships needing inspection
waters. However, development of such UUVs imposes
3 safety concerns
extreme design specifications for the hardware (Uhrich and
Watson, 1992), incurring an unaffordable cost for most labs. 4 low throughput
By contrast, UUVs for shallow waters recently have gained
5 unsustainable working time underwater to do a
more attention because of their potentially wide use
thorough job.
combined with affordable cost. At the University of
Canterbury, a UUV prototype has been recently designed To reduce the working load of divers and significantly
with the primary purpose for inspecting and cleaning the sea accelerate inspection and/or treatment, it would be highly
chests of ships [Figure 1(a)], an application with significant desirable and efficient to deploy affordable UUVs to inspect
impact in the area of bio-security. and clean these ship sea chests.
Sea chests are the intake areas in the hulls of ships for For a UUV to work properly in this scenario, the key
seawater used for ballast, engine-cooling and fire-fighting. specifications include:
196 W.H. Wang et al.

1 low cost so that it is affordable for mass deployment for • operating temperature ranges
shallow water tasks no deeper than 20m
• structural integrity for additions and tapings
2 slightly positive buoyancy so that it does not sink to the
• impact conditions
sea/water bed when power runs off
• water permeability
3 compact structure for easy on-site maintenance and
trouble shooting • visual appeal and aesthetics
4 extendable space for carrying payload to inspect and • accessibility
clean the sea chests
• versatility
5 manual or autonomous navigation at a cruise speed of
~1 m/s. • practicality

To meet the specifications, this paper presents a low cost • restrictions for future additions
UUV prototype emphasising the unique design issues and • size requirements
solutions developed for this task, as well as those attributes
that are generalisable to similar systems. Control and • corrosion and chemical resistance.
navigation are being implemented and are thus not covered The foremost design decision is the shape of the hull.
here. Inspired by torpedoes and submarines, a cylindrical hull has
been selected. A cylinder has favourable geometry for both
pressure (no obvious stress concentrations) and dynamic
2 Hull design reasons (minimum drag). To make the hull, three easily
Figure 2 shows the UUV prototype (weighing 112 kg, accessible materials were compared. The first option is to
positively buoyant), which consists of basic components, use a section of highly available PVC storm water pipe. The
including main hull, two horizontal propellers, four vertical second option involves having a hull made from a
thrusters, two batteries, an external frame and electronics composite material, such as carbon fibre or fibre glass.
inside the main hull. This section focuses on the hull design. Mandrel spinning of such a hull will allow more freedom in
radial dimensions. The process can in fact incorporate a
Figure 2 The hull structure of the vehicle, (a) top view (b) side varying radius along the length resulting in a slender,
view (c) isometric view (d) real picture of the traditional hull. However, this process requires a large
in-house made vehicle (see online version for colours) amount of design and set up time. A less desirable third
option is to use a section of metal pipe, which is prone to
corrosion and has a high weight and cost. As a result, the
PVC storm water pipe option was selected.
Two caps were designed to complete the hull, and are
attached to each end of the pipe such that they reliably seal
the hull. The caps also allow access to the interior for easy
repair and maintenance in the field.
The end cap design incorporates an aluminium ring that
is permanently fixed to the hull and a removable aluminium
plug. The plug fits snugly into the aluminium ring. Sealing
is achieved with commercially available O-rings. Sealing
directly to the PVC hull would have been more desirable,
however this option was not taken for two main reasons.
First, PVC does not provide a sealing surface as smooth and
even as aluminium and is extremely hard to machine in this
case due to the size of the pipe. Second, the PVC pipe is not
perfectly round and subject to significant variability, which
would have made any machined aluminium cap subject to
Note: (a)–(c) design drawings of the vehicle.
poor fit and potential leakage, decreasing reliability.
There are a number of different ways in which hull design The design choices made can thus better manage these
can be approached (Allmendinger, 1990). These different issues. More specifically, the design is based on self-sealing
design methods are typically specific to the situation/task. where greater outside pressures enforce greater connection
The main hull must be able to meet a number of key between the cap, seals and PVC hull portion. The O-ring
challenges. seal employed is made of nitrile, which is resistant to both
Aspects that must be considered during hull design fresh and salt water.
include:
• pressure and/or depth required
 Design of low-cost unmanned underwater vehicle for shallow waters 197

3 Propulsion and steering 4 Ballast and depth control

The design incorporates two horizontal thrusters mounted In the design, both static and dynamic diving principles are
on both sides of the UUV to provide both forward and considered. Based on these criteria four initial concepts are
backward movement. Yaw is provided by operating the developed (Wolf, 2003):
thrusters in opposing directions. The thrusters are 12V dive
1 a piston type ballast tank
scooters (Pu Tuo Hai Qiang Ltd., China) that have a
working depth of up to 20m. 2 a hydraulic pump based ballast system
The dive scooters are lightly modified to enable simple
3 an air compressor based system
attachment to the external frame of the UUV. The thruster
mounts consist of two aluminium blocks, which, when 4 direct thrust systems.
bolted together, clamp a plastic tab on each thruster. These
The first three concepts come from static diving technology,
clamps provide a strong, secure mount that can be easily
while the last concept is a dynamic diving technology.
removed or adapted to other specifications.
The piston ballast tank is one of the most common static
The force that can be generated by the thruster is
diving methods applied in submarine modelling. A piston
characterised, as shown in Figure 3. The significant linearity
ballast tank consists of a cylinder and a movable piston, and
between the thruster force and the applied duty cycle will
it works as a large syringe pump. With one end of the
significantly facilitate the design and implementation of any
cylinder connected to surrounding water, movement of the
control scheme.
piston sucks water in or pushes it out. When water fills the
Figure 3 Calibration of the motor: force with respect to duty tank, negative buoyancy is achieved, so the UUV starts to
cycle (see online version for colours) descend. Conversely, when the tank is emptied, the UUV is
positively buoyant, so it ascends. This setup also allows
control of pitch motions of the UUV. Moreover, the pistons
can be moved by a linear actuator, which is electrically easy
to control. Hence, accurate depth control can be achieved
with proper, yet straightforward, programming.
A hydraulic pumping system is similar to the piston
ballast tank, but uses an internal reservoir of hydraulic fluid
and a pump to actuate the piston’s linear motion. Control of
the valves and the pump for the hydraulic fluid allows it to
flow in and out of the cylinders, so the surrounding water
can be pumped in and out. Consequently, buoyancy of the
UUV is changed.
Air compressor systems are commonly used in some
classes of submarine. The system is composed of a storage
tank of compressed air, a water tank and two valves that are
normally closed. To descend, the vent valve is opened, so
Figure 4 Drag force of the UUV with different velocities the pressure difference results in water flowing in from the
(see online version for colours)
opening in the bottom of the water tank. When a desired
amount of water is obtained for ballast, the vent valve is
closed. In order to force the water out, the blow valve is
opened to allow the compressed air into the tank so that
water is pushed out via the bottom opening. Thus, by letting
the water in and out of the water tank, the buoyancy of the
UUV is changed.
Thrusters are a dynamic diving method. They require the
UUV to be near neutrally buoyant. This approach uses the
vertically mounted thrusters to force the UUV to dive.
Turning off the thrusters or using them at a thrust less than
the positive buoyancy allows controlled ascent. However,
A fluid drag force model is established to evaluate the speed this method consumes a lot of power to keep the UUV under
that the UUV can achieve. Figure 4 shows the relationship water, as the thrusters must remain powered at virtually all
between the drag force with respect to the relative velocity times. Being positively buoyant, however, this method is
of the vehicle. Under the full load of the two thrusters, the intrinsically failsafe, as the vehicle will come to the surface
vehicle is able to achieve a maximum forward or backward in the event of a power failure.
speed of 1.4 m/s (~5 km/hr). Selection of a suitable ballast system is dependent on
various factors, such as design specifications, size and
geometry of the UUV hull, depth required, and cost. In this
198 W.H. Wang et al.

design, the hull is made of a PVC pipe with an outer Finally, each concept has its own advantages and
diameter of 400 mm and a length of 800 mm. The required disadvantages. Comparisons are summarised in Table 1. In
working depth is 20m. Hence, the ballast system selected this design, the major driving factors for the selection of
not only has to meet the basic requirements enumerated ballast system are the cost and reliability. Piston ballast tank
above, but must also be able to fit in the hull. Preferably, all and thruster systems are reliable since these two depth
are at a relatively low cost. control methods have been widely employed in most UUV
First, installing two 160 mm inner diameter ballast tanks development. Considering the cost, the thruster system is
of 250 mm length provides a net force of ±5 kg. more effective. Hence, the thruster system is chosen as the
Additionally, the force required to actuate the piston head at final design.
20m is calculated to be approximately 6,000N. To generate
such force on the piston head, a powerful linear actuator is Table 1 Ballast comparison
needed. The specific linear actuator (LA36 24V DC input,
Diving Installation Bouyancy
6,800N max load, 250 mm stoke length) can be sourced tech
from Linak Ltd. in New Zealand. However, the linear
actuator has a duty cycle of 20% at max, which means that Piston ballast Static Internal +ve, –ve,
tanks neutral
for every 20s continuous work, it must remain off for 80s
before operating again, allowing the UUV to float Hydraulic Static Internal +ve, –ve,
pumping system neutral
uncontrolled. In addition, the cost of one linear actuator is
US$1,036, which would imply that similar actuators with Air compressor Static Internal +ve, –ve,
neutral
longer duty cycles would cost a larger amount at this time.
Taking the second option, a hydraulic pumping system Thrusters Dynamic External +ve
can be customised from Scarlett Hydraulics Ltd, Overall
Sealing Reliability
New Zealand. The overall system has dimensions of cost (US$)†
500 mm × 250 mm × 250 mm. It consists of a 1.2 KW DC Piston ballast Difficult Used in most 2,500
motor, a pump, a 4L hydraulic fluid tank, two dual solenoid tanks remote
valves and two cylindrical tanks. This system meets the submarines
required specifications, but has some drawbacks. In Hydraulic Difficult Not reliable 2,710
particular, it occupies too internal space of the hull and pumping system
weighs approximately 20 kg (a significant addition of Air compressor Difficult Air on board is 420
weight). In this case, the overall hydraulic pumping system limited,
will cost up to approximately US$2,264. compressed air
The third option air compressor system is cost effective hard to handle
and is easy to operate by controlling the vent and blow Thrusters None Used in most 500
valves. However, the lack of accuracy in controlling UUVs with big
compressed gas is a major disadvantage. In addition, size
performance and operating time are limited by the amount Note: †The cost is estimated as an overall system.
of stored gas. In this design, a 10L tank would be needed to
There are four thrusters vertically mounted around the UUV
fulfil the changes in buoyancy. In other words, a gas
with one at each corner (see Figure 2). Mounting four
cylinder containing 10L of air compressed to at least three
thrusters produces a total of 20 kg thrust force at full load
bars is required for a single diving and rising cycle. Hence,
and allows a wide range of motion control. They enable the
to refill the gas cylinder, the UUV must float to the waters
control of not only the vertical up and down motion, but
surface before all the air runs out or risk being lost.
pitch and roll motions. To achieve this control, each thruster
Regarding the on-site requirement, that the UUV should
is connected to a speed control module that can be
operate for hours, the air tank must either be much bigger or
controlled via a central microprocessor. By inputting
far more highly pressurise, which leads to safety issues.
different digital signals, various forces thus speeds are
The fourth option thrusters are different from the
generated. Therefore, desired motion control can be
previous three systems that all had to be installed inside the
obtained by different combinations. For example, two of the
UUV. In contrast, thrusters can be attached externally.
four vertically mounted thrusters on the same side of the
Hence, sealing is not as critical as it is for the other
vehicle can be turned on to achieve roll motion. Likely, two
concepts. If the vehicle is trimmed positively buoyant, it is
of the four vertically mounted thrusters on the same end
also reasonably fail-safe, unlike the other three methods.
(near the front or end cap) of the vehicle can be turned on to
Additionally, the thrusters can be sourced from Pu Tuo Hai
achieve pitch motion.
Qiang Ltd., Zhou Shan, China for US$55/unit, a reduction
of 12–20 times in cost if two are used. Each thruster fits in a
215 mm × 215 mm × 80 mm box and is driven by a 12V DC
motor with a max thrust force of 5 kg under water. By
mounting the desired number of thrusters, a wide range of
motions can be controlled, such as pitch and roll control.
 Design of low-cost unmanned underwater vehicle for shallow waters 199

5 Electronics and control power than either of the prior two options, it provides
immense processing power, memory and a diverse range of
5.1 Power supply peripherals. It is therefore chosen in this initial design for
For long term operation, this design must locate the power the following primary reasons:
supply on-board, unlike many current models that receive • added power requirements were not an issue since we
power over an umbilical link (Chardard and Copros, 2002). have a sizeable power supplies
Since all the systems onboard the UUV are electric, sealed
lead acid batteries are chosen for the power supply. These • processing power is more than adequate for this initial
batteries have high capacity and can deliver higher currents, design and future developments
than other types of rechargeable battery (Schubak and Scott,
• large volumes of memory are available, both volatile
1995; Bradley et al., 2001). They are stable, inexpensive,
for program execution and solid state for storage of
mechanically robust and can work in any orientation, all of
gathered data
which are important considerations in a vehicle of this type.
To supply enough current for the entire machine several • despite not having direct access to sensors and control
batteries have to be joined together. Instead of adding dead units, a diverse range of peripherals available can be
weight to achieve neutral buoyancy extra batteries can be used, including Universal Serial Bus (USB), RS232 and
added as needed so that the total operating time of the UUV Ethernet, enabling a potentially greater range of sensors
is higher than that required for a given application. and sensor platforms for developing broad ranges of
It is also highly desirable to locate the battery specific applications
compartments separate from the main hull so that they can
• a USB module is already provided for a webcam for
be interchanged in the field without opening the sealed main
initial image sensing applications and an Ethernet
hull. To accommodate this requirement two tubes are fitted
module is provided for remote connection.
below the hull to house batteries. Within these tubes the
batteries are connected to two bus bars. Each battery is Figure 5 On board computer (see online version for colours)
fused prior to connecting to the bus bar, and the bars are
isolated to the greatest extent possible to increase safety.
These bus bars are then wired into the main hull, where a
waterproof socket enables the quick interchange of battery
compartments. A similar bus system exists inside the hull
with connections to motors and electronic power supplies.
Each of these internal connections is similarly fused. Longer
term, it would be desirable to intelligently monitor the bus
to track the state of each battery and overall power
consumption.

5.2 Central processing unit

The central processing unit is responsible for accessing
sensors, processing data and setting control outputs such as An AMD Sempron 3000+ processor and ASUS M2N-PV
motor speeds. Several systems are considered for motherboard are used for this purpose. These models have
this unit, an embedded system using microprocessors, lower power requirements and heat generation. Software
field-programmable gate array (FPGAs) or a small desktop interfaces this unit with sensors and motor controllers, as
PC. A microprocessor system, most likely based on an well as to a remote control PC. An automotive power supply
Advanced RISC Machine (ARM) processor would have low (Exide, Auckland, NZ) is used to provide power for the
cost, size and power requirements and is easy to interface to computer. It takes a 12V DC input and converts it to the
both analogue and digital sensors, motors and other Advanced Technology Extended (ATX) standard power
actuators. The processing power and memory allocations of supply required by the PC. This module is also designed to
these microprocessors are all more than sufficient for the be used in an electrically noisy and hostile environment and
simple control tasks likely to be required, but would struggle is ideally suited the specific design situations considered.
with larger sensor or processing tasks, such as image
processing. An FPGA system would also be small and have 5.3 Sensors
low power requirements, but would be more expensive.
While FPGAs work very well for fast, complex processing When the UUV is used autonomously, after development
tasks such as image processing, their complexity in design there will be a large and extensive sensor suite onboard.
and programming necessitates their use in parallel with Currently, the sensors onboard measure:
other more flexible CPU choices. The last system • water pressure, from which depth can be determined
considered is a small desktop PC (Figure 5). Although a
desktop PC is bigger, more expensive and consumes more • water temperature, inner hull temperature and humidity
200 W.H. Wang et al.

• the UUV position in the three principal axes: yaw, pitch Mandeno Electronics for US$121. This sensor measures up
and roll to twice the depth required and outputs an analogue output
between 0 mV and 100 mV. Thermocouples from Farnell
• visual or digital image feedback via a webcam.
Electronics (Christchurch, New Zealand) are used to
Submersible pressure sensors that are salt water tolerant and measure the water temperature and provide an analogue
can measure up to the pressures required are difficult to output relative to the temperature difference between the
acquire at low cost. The sensor chosen was sourced from two ends of the thermocouple.

Figure 6 The block diagram for electronic systems and control

For navigation and control of the UUV, normally an

TMP100 sensors (Texas Instruments) are used to measure
inertial measurement unit (IMU) is used to give the
the base temperature of the thermocouple and the hulls
instantaneous position. An IMU qualified for this job costs
interior temperature. These sensors give a digital output
around US$20,000. To reduce the cost, an alternative
using the I2C protocol. A HF3223 humidity sensor
option, which will be integrated to the system in future
(Digi-Key) is used to measure humidity inside the hull. A
development, is stereo image guidance. Two cameras can be
MMA7260QT accelerometer (Freescale Semiconductor) is
employed onboard to reconstruct the 3-D mapping of the
used to calculate orientation. The accelerometer has a 0V to
environment for applications such as object detection and
2.5V analogue output. The connection of the sensors is
autonomous UUV navigation.
shown in Figure 6.
To eliminate signal noise, An Atmel AT90USB82
microprocessor is connected to the USB ports of the 5.4 Propulsion motor driver
computer to move all noise sensitive data to the acquisition For the six motors (two for horizontal propulsion and four
points. The analogue sensors are amplified using an for vertical ballast control), three RoboteQ AX2500
INA2322 instrumentation amplifier, if necessary, and read (Figure 7) motor controllers are used for control. Each
by an ADS7828 analogue to digital converter. This controller is able to control two motors up to 120 amps,
converter is then connected to the Atmel microprocessor much higher than the 25 amps needed by the motors
using a common I2C bus with the TMP100. The humidity selected. The controllers are controlled via RS232 (serial
sensor is attached to a clock input which converts the port) interfaces, which are already available on the
frequency based signal to a humidity based reading. The computer motherboard. Computer control of the controllers
microprocessor performs some basic processing on this is easily achieved through a LabView or MATLAB
data, temperature compensating the pressure sensor and interface, either manually or automatically, where both
thermocouple and calculating yaw, pitch and roll from the interfaces have been implemented to allow greater user ease
accelerometer readings. of use.
Vidual or digital image sensing is included via a
Logitech webcam connected directly to the on-board
computers USB port. The video stream can be sent back
over a wireless remote control network connection to the
remote PC. At this stage, no image processing is done on
this stream on-board and it is included purely to assist in
manual control of the UUV at this time and for use in later
application development.
 Design of low-cost unmanned underwater vehicle for shallow waters 201

Figure 7 The controller for both horizontal and vertical thrusters References
(see online version for colours)
Allmendinger, E. (1990) Submersible Vehicle Systems Design,
Society of Naval Architects and Maringe Engineers, Jersey
City, NJ.
Ballard, R. (1987) The Discovery of the Titanic, Warner/Madison
Press Books, NY, NY.
Bradley, A., Feezor, M., Singh, H. and Sorrell, F. (2001) ‘Power
systems for autonomous underwater vehicles’, IEEE Journal
of Oceanic Engineering, Vol. 26, No. 4, pp.526–538.
Caccia, M. (2006) ‘Autonomous surface craft: prototypes and
basic research issues’, 14th Mediterranean Conference on
Control and Automation, Italy, June.
Chardard, Y. and Copros, T. (2002) ‘Swimmer: final sea
demonstration of this innovative hybrid AUV/ROV system’,
2002 International Symposium on Underwater Technology,
5.5 Control system and communications Japan, April.
Curtin, T. and Bellingham, J. (2001) ‘Autonomous ocean-sampling
During testing and development, remote control is required networks’, IEEE Journal of Oceanic Engineering, Vol. 26,
for the UUV. Sensors readings need to be sent to a user and pp.421–423.
control signals sent back to the UUV. Displaying the video Griffiths, G. and Edwards, I. (2003) ‘AUVs: designing and
feed from the webcam is also desired to provide the operator operating next generation vehicles’, Elsevier Oceanography
with visual feedback. High frequency radio transmissions Series, Vol. 69, pp.229–236.
are impossible underwater due to the high losses Horgan, J. and Toal, D. (2006) ‘Review of machine vision
encountered during the air/water boundary (Leonessa et al., applications in unmanned underwater vehicles’, 9th
2003). Lower frequency transmissions could have been used International Conference on Control, Automation, Robotics
to communicate with the UUV, but they do not possess and Vision, Singapore, December.
enough bandwidth to send the required data. An umbilical Kondoa, H. and Ura, T. (2004) ‘Navigation of an AUV for
Ethernet cable is being used for this remote link between the investigation of underwater structures’, Control Engineering
Practice, Vol. 12, pp.1551–1559.
UUV and an external control computer for this development
phase. Figure 6 shows the electronics and control structure. Leonessa, A., Mandello, J., Morel, Y. and Vidal, M. (2003)
‘Design of a small, multi-purpose, autonomous surface
Note that in an actual, developed application or final
vessel’, OCEANS 2003, USA, September.
development thereof, the robot will be acting autonomously
Lygouras, J., Lalakos, K. and Tsalides, P. (1998) ‘THETIS: an
and this umbilical will not be required.
underwater remotely operated vehicle for water pollution
measurements’, Microprocessors and Microsystems, Vol. 22,
No. 5, pp.227–237.
6 Conclusions Rife, J. and Rock, S.M. (2002) ‘Field experiments in the control of
a jellyfish tracking ROV’, MTS/IEEE Oceans’02, USA,
A UUV has been successfully designed and prototyped for
October.
shallow water tasks, such as inspecting and cleaning sea
Schubak, G. and Scott, D. (1995) ‘A techno-economic comparison
chests of ships. It features low cost and wide potential use of power systems for autonomous underwater vehicles’, IEEE
for normal shallow water tasks with a working depth up to Journal of Oceanic Engineering, Vol. 20, No. 1, pp.94–100.
20m and a forward/backward speed up to 1.4 m/s. Each part Smallwood, D. and Whitcomb, A.L. (2004) ‘Model-based dynamic
of the UUV is deliberately chosen based on a comparison of positioning of underwater robotic vehicles: theory and
readily available low cost options when possible. The experiment’, IEEE Journal of Oceanic Engineering, January,
prototype has a complete set of components including Vol. 29, No. 1, pp.169–186.
vehicle hull, propulsion, depth control, sensors and Smallwood, D., Bachmayer, R. and Whitecomd, L. (1999) ‘A new
electronics, batteries and communications. The total cost for remotely operated underwater vehicle for dynamics and
a one-off prototype is less than US$10,000. With these control research’, Proc. UUST’99, USA, August.
elements, a full range of horizontal, vertical and rotational Tivey, M., Johnson, H., Bradley, A. and Yoerger, D. (1998)
control of the UUV is possible including computer vision ‘Thickness of a submarine lava flow determined from
sensing. The overall underwater vehicle will be a good near-bottom magnetic field mapping by autonomous
underwater vehicle’, Geophysical Research Letters, Vol. 25,
platform for research, as well as for its specific applications, pp.805–808.
many of which are growing in importance like the sea chest
Uhrich, R. and Watson, S. (1992) ‘Deep-ocean search and
inspection case noted here. inspection: advanced unmanned search system (AUSS)
concept of operation’, Naval Command, Control and Ocean
Surveillance Center, RDT&E Division, Tech. Rep. NRaD TR
1530, San Diego, CA, USA, November.
202 W.H. Wang et al.

Wernli, R. (2001) ‘Low cost UUVs for military applications: is the

technology ready?’, Pacific Congress on Marine Science and
Technology 2001, USA, July.
Willcox, S., Vaganay, J., Grieve, R. and Rish, J. (2001) ‘The
bluefin BPAUV: an organic wide-area bottom mapping and
mine-hunting vehicle’, Proc. UUST’01, USA, August.
Williams, C. (2004) ‘AUV systems research at the NRCIOT: an
update’, 2004 International Symposium on Underwater
Technology, Taiwan, April.
Wolf, M. (2003) ‘The design of a pneumatic system for a small
scale remotely operated vehicle’, Bachelor’s thesis, MIT,
May.